Work of Breathing After Extubation

Work of Breathing After Extubation* Abraham M. Ishaaya, MD; Steven D. Nathan, MD, FCCP; and Michael ]. Belman, MD, FCCP Recently we showed that work of breathing was higher in the immediate period after extubation as compared with spontaneous breathing through an endotracheal tube. In this study, we evaluated the glottis and trachea as potential sites of increased airway resistance after extubation. We measured breathing pattern, work of breathing, and pressure time product in eight patients during weaning from mechanical ventilation. We acquired data during pressure support ventilation and spontaneous breathing via the ventilator, with the endotracheal tube in place, and after extubation. During bronchoscopy at the time of extubation, we examined the trachea and measured the cross-sectional area of the glottis. Work of breathing and pressure time product were significantly lower during pressure support ventilation as compared with spontaneous breathing after extubation (0.43±0.10 vs 1.49±0.10 J/L and 101 ±22 vs 299±30 em H 20-s/min, respectively; p<0.05). However, both indexes were significantly higher after extubation as compared with breathing through the endotracheal tube (1.49±0.10 vs 0.95±0.12 J/L, 299±31 vs 196 ± 26 em H20 · s/min respectively; p<0.05). During bronchoscopy, no tracheal or glottic narrowing was detected. The glottic cross-sectional area was successfully

Jnwork a recently completed study, we showed that the of breathing (WI) was significantly in1

creased in the immediate postextubation (EXTO) period as compared with the WI while breathing spontaneously through an endotracheal tube (ETT). This was an unexpected result, since it is generally believed that when inspiratory flow rates are comparable, the added resistance of an ETT increases Wr. 2•3 Both the increased length and the smaller internal diameter of an ETT (50 to 64 mm 2 for 8- and 9-mm tubes, respectively) as compared with the normal glottis are thought to be responsible for the increased ventilatory load. The goal of this study was to examine if the increased WI in the postextubation period could be caused by changes in the trachea or glottis because previous studies have documented that endotracheal intubation causes upper airway damage. 4·5 MATERIALS AND METHODS

Clinical Procedures Eight patients, identified as eligible for we
204

measured in four patients at the onset of inspiration and found to be 140 ± 15 mm 2. This value was larger than the mean cross-sectional area of the endotracheal tubes used in these patients (50 mm 2). We conclude that neither tracheal nor laryngeal disease caused the increase in work of breathing after extubation. Our data suggest that upper airway narrowing at a more proximal site, such as the oropharynx or velopharynx may be the cause of the increase in respiratory work. (Chest 1995; 107:204-09)

Ag=glottic area; APg=anteroposterior diameter of glottis; ETT=endotracheal tube; EXT=post-extubation; EXTO= immediately EXT; EXT2=2-h EXT; EXT24=24 h EXT; £=breathing frequency; Paw=airway pressure; PEEPi= intrinsic positive end-expiratory pressure; Pes=esophageal pressure; PS=pressure support; PSmin =minimal PS; PSV=PS ventilatio~?; PSVO=PSV of 0 em H 20; PTP= pressure time product; V=flow; VT=tidal volume; WI= inspiratory work of breathing

Key words: intubation; mechanical ventilation; respiratory failure; upper airway; work of breathing

mary physicians, were studied during weaning (Table 1). There were five men and three women. The mean age of the patients was 71 (range, 23 to 91) and the mean time of intubation was 5.5 days (range, 1 to 14). Three of the patients had congestive heart failure; two had COPD and pneumonia, and one of the patients was intubated because of status epilepticus. All patients were ventilated by means of a Puritan Bennett 7200A Ventilator. At the time of weaning, all patients were alert and cooperative. The experimental protocol was approved by the institutional review board of the hospital, and all patients gave their informed consent prior to participation in the study. A level of pressure support ventilation (PSV) was calculated for each patient using the formula: minimal pressure support (PSmin)=PIFR X R, where PIFR is the peak spontaneous inspiratory flow rate and R is the total resistance of the respiratory system. 6 Although the formula for PSmin has not been validated, 1 it serves as a useful starting point to standardize the initial pressure support (PS). The PSmin has been defined as that level of PS which is necessary to simulate spontaneous breathing. At PSmin, it is assumed that the PS is just adequate to overcome the resistance of the ETT and the ventilator circuit. 6 Peak spontaneous inspiratory flow rate was measured during two spontaneous breaths while the patient was receiving ventilator assistance (constant positive airway pressure, 0 em H 20) and the average value used. The R was calculated from measurements made during a controlled breath using the formula: R=Ppeak-Pplateau/ mean inspiratory flow rate 6 Prior to extubation, an esophageal catheter with a 10-cm balloon at the distal end (part No. 700-3-100 Bicore, Irvine, Calif) was passed transnasally and positioned in the lower third of the esophagus. The position of the Work of Breathing After Extubation (lshaaya, Nathan, Belman)

Table !-Patient Demographics Patient

Age/ Sex

ETT Diameter

o. Ventilator Days

PSmin, em HzO

2 3 4 5

89/ M 91 / F 68/ M 90/ F 32/ M

7.5 6.0 8.0 7.0 8.0

7 3 2 14 1

6 5 9 7 4

6 7 8

87/ F 78 / M 23/ M

7.5 8.0 8.0

2 3 12

6 5 6

Diagnosis* COPD CHF COPD CHF Airway Protection CHF Pneumonia Pneumonia

*CHF=congestive heart failure. tube was checked by means of the airway occlusion technique.7 Flow and proximal airway pressure were recorded from a pneumotachograph/ pressure sensor (part No. 700-2-300 Bicore, Irvine, Calif) placed between the ETT and the Y piece of the breathing circuit. Both the flow and pressure sensors were calibrated prior to use in each patient. The pressure and flow signals were transmitted to a Bicore CPlOO monitor. The performance of the Bicore pressure transducers and the pneumotachograph have been examined and found to be accurate. 1 The following indices were recorded: tidal volume (VT), respiratory frequency (f), proximal airway pressure (Paw), flow (V ), and esophageal pressure (Pes). Intrinsic positive end-expiratory pressure (PEEPi) was calculated as the pressure difference between the Pes at the onset of inspiratory effort, as defined by the initial reduction in end-expiratory Pes, and the Pes at the onset of inspiratory flow 8 •9 The WI of the patient, expressed as Joules per liter was calculated from the area subtended by the Pes developed during inspiration and the relaxation curve of the chest wall (estimated chest wall compliance equal to 200 mL / cm Hz0). 9 ·10 The pressure time product (PTP) was derived from the following formula: PTP= f i~Pes dt, where To is the onset of inspiratory effort and Ti is end inspiration . The chest wall static recoil pressure time curve and the PTP due to PEEPi were included, as described by Sassoon and coworkers 8 These data were digitized by the CPlOO monitor and then transmitted in real time to a personal computer for storage and subsequent analysis. Measurements were made for 2 min after the patients had developed stable breathing patterns during the PSmin and PSmin25% (the latter is defined as the minimal level of PS calculated as described previously minus 25%). In addition, the same measurements were made while the patients breathed spontaneously via the ventilator at a pressure support ventilation of 0 em H 2 0 (PSVO). All breathing modalities prior to extubation were applied in a random order. Ventilator assistance was then discontinued, and measurements were made with the patients breathing spontaneously through an ETT. Subsequent measurements were made while the patients breathed spon taneously through a wide-bore mouthpiece immediately postextubation (EXTO), 2 h EXT (EXT2), and 24 h EXT (EXT24). All measurements were made with the patients in the semirecumbent position. We performed bronchoscopy at the time of extubation. Prior to extubation, the trachea and pharynx were anesthetized with 2% lidocaine, but no sedation was given. The bronchoscope was passed through the ETT which was then withdrawn in order to allow inspection of the trachea, larynx, and supraglottic region. With a video camera attached to the bronchoscope, we recorded the appearance of the airway. We used a bronchoscope brush to measure the distance between the tip of the bronchoscope and the vocal cords, and videoscopic recordings were made at a known distance above the cords. We calculated the real anteror-posterior diameter of the glottis (APg) from a measurement of the APg on

the screen of the videomonitor (screen APg) utilizing a magnification factor derived from the distance of the bronchoscope from the cords. The real glottic area (Ag) was then calculated from the following equation: Real Ag=screen Ag X (real APg 2 / screen APg 2, where screen APg 2 =the area of the glottis on the videoscreenll Statistical Analysis

The mean data for each ventilatory mode were calculated from the breaths acquired during each 2-m in measurement period. We used an analysis of variance for repeated measures in order to compare the various indexes. Where significant changes were found , post hoc testing between groups was done using the Newman Keuls test. The results are expressed as mean ± SEM. Before and after values for paired samples were compared using the Student's t test (Crunch, Crunch Software, Oakland, Calif). A probability value of less than 0.05 was considered significant. In order to assess the agreement in WI between EXTO and PSVO, we compared the Wr during these two modes using the technique described by Bland and AltmanJ 2 RESULTS

Recordings from an individual patient are shown in Figure 1, where tracings of flow, volume, Paw, and Pes during ETT and EXTO are shown. The increase in Pes despite similar VT values, is apparent. The mean VT. f, PEEPi, and delta Pes (the difference between Paw and Pes) of all patients are shown in Table 2. No significant changes were noted in these variables. The mean PSmin was 6.0 em H 20 with a range of 4 to 9 em H 20. During EXTO, the mean WI was significantly higher than the mean WI during ETT (p<0.01 [Fig 2]). The WI during PSmin was significantly lower than the WI during PSVO, ETT, and EXTO (p <0.05 [Fig 2]). The WI values during both PSmin and PSmin-25% were significantly lower than the WI during EXTO (p<0.05). Thus, both pressure support ventilation modes overestimated the amount of support required to overcome the resistance of the ETT and ventilator circuitry as shown by the lower WI during these modes as compared with the WI during EXTO. The mean WI during PSVO and EXTO were not significantly different (1.49 ± 10 J/ L vs 1.19 ± 15 J/ L, and in Figure 3 the mean bias of WI during EXTO vs PSVO (+0.3 J/ L) is shown . Although CHEST / 107 / 1 I JANUARY, 1995

205

ETT

EXTO

J-. -~ ,;_;./

v

""

1-. ··-

... .:-:-:.., ..

/ .."'

~p

...

v -~ k

v•

IC

v... --~ ~ V--- ·---" ~../. ........... /

/

. ·.:...

~

II~

.

~r

FIGURE l. The tracings show V, VT, Paw, and Pes during spontaneous breathing through an ETT (left pane!) and EXTO (right panel) in one patient. During EXTO, there is a larger change in Pes for a sim-

ilar V and VT.

the mean values were reasonably close, the individual differences show a wide scatter and thus poor precision. Thus prediction of Wr during EXTO from measurement of Wr during PSVO is unreliable. Figure 4 shows the individual changes in Wr for the eight patients. In only one patient did the Wr decrease postextubation. Two patients were reintubated because of respiratory distress. One patient required reintubation after approximately 1 hand another after 6 h. Neither patient was monitored by the Bicore device during the period of decompensation. The remaining six patients were successfully weaned from mechanical ventilation. The measurements of Wr and PTP made during the first 24 h postextubation, showed that the mean WI and PTP EXTO, were not significantly different from EXT2 and EXT24 h (1.52 ±0.12, 1.81 ±0.07, 1.32 ±0.3 J/ L and 295 ±35, 404±93, 246±33 em HzO/ min, respectively) . The PEEPi at ETT and EXTO were not significantly different (1.9 ± 0.2 and 3.2 ± 0.7 em HzO, respectively). The end-expiratory Pes during ETT and EXTO were similar (9.9 ± 1.4 em HzO and 10.2 ± 1.2 em H 20 , respectively).

The trachea and glottis were visualized in all patients, and apart from generalized erythema in the trachea and glottis no narrowing was observed. In none of the patients were significant secretions noted. In four of the eight patients, we were able to position the bronchoscope in a stable position in order to make recordings adequate for subsequent measurements of Ag. In these patients, the mean cross-sectional area of the laryngeal aperture was 140 ± 15 mm 2 at the onset of inspiration as compared with the mean area of 50mm 2 for the ETT (Table 3). DISCUSSION

We found that the mean WI and PTP were significantly higher EXT as compared with spontaneous breathing through an ETT. These changes persisted for at least EXT24. During bronchoscopy, we did not find abnormalities in the trachea and glottis which could account for these changes. Abnormalities of the upper airway proximal to the glottis could be responsible for the increase in WI postextubation, but this region was not specifically examined. The local anesthesia that was administered at the

Table 2-Measurements During Modes of Weaning in All Patients*

PS PS-25% PSVO Breathing with ETT EXTO

VT, 1

f breaths per minute

PEEPi, em HzO

Delta Pes, em HzOf

0.39±0.03 0.39± 0.03 0.38±0.03 0.36±0.03 0.28±0.06

21±1 21±1 22±2 22±2 21±2

2.3±0.4 3.5±0.7 3.1 ±0.4 1.9±0.2 3.0±0.7

10.4 ± 1.3 13.1 ± 1.4 17.8 ± 1.9 14.7±1.6 20.3±0.8

*Values are mean ± SEM. Abbreviations: VT=tidal volume; F = respiratory frequency; PEEPi=intrinsic PEEP; delta Pes =difference between airway pressure and esophageal pressure. fDelta Pes is the difference between Paw and Pes.

206

Work of Breathing After Extubation (lshaaya, Nathan, Belman)

2.0

-.........,

D

w,

n

PTP

=8

-

350 300

_J

250 c:

E

200 Ui

....1

.........

0N

150 ::r:

~

100

so 0.0

1.5

'....,

......., 0

> (/) D..

0 1-

D.

w

1D.

0.5

I

E

u

•

1.0

•

•

•

•

•

X

0.0

~-

•

.... 0

PS(mln) PS-25 PSVO

Err

EXTO

2. The mean WI (open column) and PTP (solid column) during the different modes of breathing are shown. The error bars represent the SEM. FIGURE

time of bronchoscopy was unlikely to be responsible for the increase in WI. In our prior study, in which no local anesthesia was used, we showed a similar increase in WI postextubation. 1 Furthermore, the increase in the WI persisted for 24 h, longer than would be expected if the effect was due solely to the local anesthetic. Although the presence of secretions in the upper airways might cause partial airway obstruction,13 we believe that this is an unlikely source of the increased work, since no patient was noted to have excess secretions at the time of bronchoscopy. Moreover, the patients were suctioned at the time of the bronchoscopy, and the initial recordings were taken immediately thereafter. The initial measurements at EXTO might have been influenced by the act of extubation plus the fact that the measurements were made immediately after the bronchoscopy. However, similar results were obtained in our previous study 1 in which no bronchoscopy was performed. Moreover, the increases in WI persisted and were present at EXT2 and EXT24 has well as well as after bronchoscopy. The increase in WI and PTP could arise from changes in airway resistance (either upper or lower airways) or changes in chest wall or pulmonary compliance. However, we did not find clinical evidence of bronchospasm, airway secretions, or pulmonary edema, conditions which could rapidly change pulmonary impedance and thus increase WI postextubation. A marked increase in PEEPi postextubation also could increase WI by acting as a threshold load prior to the onset of inspiratory flow. 9.l 0 The change in PEEPi was not significant, and the Pes at end-expiration during spontaneous breathing through an ETT and EXTO was essentially unchanged, evidence against an increase in end-expiratory lung volume which would be expected if PEEPi were to rise substantially. Thus, by a process

G)

"0

-0.5

-1.0 1.0

• 1.2

1.4

1.6

1.8

2.0

AVERAGE W1 {EXTO, PSVO) FIGURE 3. The difference in WI between EXTO and PSVO is shown for each patient on the ordinate, while the abcissa shows the average WI. The middle horizontal solid line represents the mean difference (bias) in WI (+0.30 J/ L), while the upper and lower horizontal lines show the 95% confidence bands (precision) for the mean difference ( ± 1.10 J/ L, from + 1.40 to -0.80 J/ L).

2.0 1.8

,....... ...J

1.6

'

1.4

~

1.0

-,

........_..

1.2

0.8 0.6 0.4

rn

EXTO

4. Shows the individual values for WI during ETT and EXTO. All patients apart from one showed an increase in WI.

FIGURE

of elimination, we believe that the upper airway is the likely source of the increase in inspiratory work. The increases in WI and PTP were similar to those found in our last study .I There are relatively few data available regarding the behavior of WI postextubation in patients after episodes of respiratory failure. While we cannot find previous reports of an increase CHEST /107/1/ JANUARY, 1995

207

Table

3-Bronchoscopy Results

Patient

Cross-Sectional Area of ETT, mm 2

Laryngeal Area at Expiration, mm 2

3 5 7 8 Average

50 50 50 50 50

115 138 136 170 140±15

in WI postextubation, the data of Brochard et aF 0 showed that when WI is expressed in joules per liter rather than the work rate in joules per minute, there is no significant decrease postextubation . Moreover, the lung and airway resistances in their study did not decrease significantly as expected postextubation. Individual data were not shown, but there was an overlap in the range of WI values of patients while breathing spontaneously through an ETT and EXT. Thus, it is conceivable that in individual cases, the Wr actually increased. Our patient sample may represent one end of the spectrum at which there is a mean increase in Wr. In making comparisons with previous studies, it is important to compare work expressed in joules per liter rather than work rate joules per minute. As shown by Fleury and coworkers,9 only in the former case is work related to pulmonary mechanics, whereas work rate (expressed as joules per minute) is sensitive to changes in minute ventilation. For example, in their study, the patient with the highest work rate had the lowest airway resistance but the highest minute ventilation. In contrast to our findings are several studies which have documented increases in work during intubation and a decrease after removal of the endotracheal tube. 2.l 4 However, these studies have used normal subjects intubated for a few hours only or subjects who placed the tip of the ETT in their mouths.3 •15 In agreement with these results, in our laboratory we have found that WI measured with the Bicore Monitor in normal subjects during spontaneous breathing through a mouthpiece was lower than WI while breathing through an ETT inserted in the mouth (unpublished observations) . Experiments which examined the behavior of airflow and resistance during in vitro experiments 16.l 7 with ETT are not applicable to our current study. Previous literature has documented the pathologic changes that occur in the trachea and larynx after intubation. 4•5 Although we frequently noted erythema and some mild swelling of the cords, we did not see obvious narrowing of the glottic and subglottic regions. In four patients, we succeeded in measuring the cross-sectional area of the glottis and found this to be larger than the mean cross-sectional area of the ETT used in these patients. In the other patients, 208

the glottis and trachea were well visualized, and no narrowing was noted. Unfortunately, because of coughing in this group, we were unable to position the bronchoscope in a stable position in order to make glottic measurements. Since we did not find evidence of airway narrowing in the regions visualized, the possibility remains that other sites are responsible for the increase in Wr. These include the velopharynx and oropharynx which were not specifically examined. In most cases, we performed the bronchoscopy through the ETT and therefore bypassed the pharynx. In the patients in whom the bronchoscope was inserted transnasally, narrowing of the pharynx may have been present but was not specifically looked for and could have been missed. Future studies to specifically examine the velo and oropharynx in the EXTO are planned. The possible role of edema or muscle hypotonia in this region in causing increased WI and failure to wean from endotracheal intubation also will be evaluated. If important, the use of continuous positive airway pressure applied via nasal or face mask may be useful. Although the use of the PSmin formula is controversial, the levels of PS corresponding to PSmin and PSmin-25% that we used are similar to those in common clinical use. As in our previous study, these levels overestimated the level of PSV required to overcome the resistance of the ETT tube and ventilator .I Therefore, our data suggest that tolerance of low levels of PS may not be a reliable predictor of successful extubation. The WI EXT, most closely approximated the WI during PSVO, but in individual patients there was poor agreement between the two modes (Fig 3). Hence, the WI during PSVO, like PSvmin, is not an accurate indicator of the WI EXTO. REFERENCES

1 Nathan SN, Ishaaya AM, Koerner SK, eta!. Prediction of pressure support during weaning from mechanical ventilation. Chest 1993; 103:1215-19 2 Kaplan JD , Schuster DP. Physiological consequences of tracheal intubation. Clin Chest Med 1991; 12:425-32 3 Gal TJ, Suratt PM. Resistance to breathing in healthy subjects following endotracheal intubation under topical anaesthesia. Anesth Analg 1980;59:270-7 4 4 Colice GL, Stukel T A ,Dain B. Laryngeal complications of endotracheal intubation and tracheotomy. Chest 1989; 96:877-84 5 Stauffer JL, Olson DE, Petty TL. Complications and consequences of endotracheal intubation and tracheotomy. Am J Med 1981; 70:665-76 6 Hughes CW, Popovich J. Uses and abuses of pressure support ventilation. J Crit Illness 1989; 4:25-32 7 Baydur A, Behrakis PK, Zin W A, et al. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982; 126:788-91 8 Sassoon CSH, Light RW, Lodia R, et al. pressure time product during continuous positive airway pressure, pressure support ventilation and T-piece weaning from mechanical ventilation. Am Rev Respir Dis 1991; 143:469-75 Work of Breathing After Extubation (lshaaya, Nathan, Belman)

9 Fleury B, Murciano D, Talamo C, eta!. Work of breathing in patients which chronic obstructive pulmonary disease in acute respiratory failure. Am Rev Respir Dis 1985; 131:822-27 10 Brochard L, Rua F, Lorino H , eta!. Inspiratory pressure support compensates for the additional work of breathing caused by the endotracheal tube. Anesth. 1991 ; 75:739-45 11 Collett PW , Engel LA. Changes in the glottic aperture during bronchial asthma. Am Rev Respir Dis 1983; 128:719-23 12 Bland JM, Altman DC. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307-10 13 Jubran A, Tobin MJ. Use of flow-volume curves in detecting

14 15 16 17

secretions in ventilator-dependent patients [abstract]. Am J Respir Crit Care Med 1994; 150:766-69 Habib MP. Physiologic implications of artificial airways. Chest 1989; 96:180-84 Fiastro JF, Habib MP, Quan SF. Pressure support compensation for inspiratory work due to endotracheal tubes and demand continuous airway pressure. Chest 1988; 93:499-505 Demers RR, Sullivan MJ, Paliotta J. Airflow resistance of endotracheal tubes [editorial]. JAMA 1977; 237:1362 Bolder AR, Healy EJ, Beatty PCW, et a!. The extra work of breathing through adult endotracheal tubes. Anesth Analg 1986; 65:853-59

CHEST / 107 / 1 / JANUARY, 1995

209